The Fifo buffer is disabled by default. I’ve tried like that as well as with it on and every buffer setting between min and max.
So it’s definitely my computer then right? Something software or hardware related?
That is a leap. It is also possible the character was not sent or something timed out on either end. 
Great… 
Sorry Ev, just didn’t want you to go off and buy a new PC thinking it was all your hardware’s fault. 
All you know is something didn’t make it from one end to the other and the program on the PC timed out and logged that. Still, that’s better than not knowing anything since it is something to work from.
I just don’t know where to go next. So we know this:
Serial is on motherboard
Works with SSC-32 and SEQ, powerpod and RIOS.
Completes download to random amounts (10%, 85%, etc.)
Recognizes the atom
For some reason, pc and Atom lose communication at seemingly random stages of programming. But never complete.
Same result with all IDE versions
Works on laptop serial port fine.
USB to serial adapter and PCI Serial Card have the exact same results.
FIFO buffer changes nothing
Port settings change nothing
No Bios settings available (only enable/disable serial port)
Tried 2001 and 2006 drivers for serial port.
Running Windows XP Media Center Edition.
The only thing I can think of that might make my computer different from most desktops, is that I’m running Window XP Media Center Edition. Maybe there is something with that, that is causing the problems.
I’m sure everyone is sick of hearing me complain about this but it’s a bit ridiculous. How can this not work? This is brand new computer. And the serial port works fine with everything else. And I’ve spent $40 trying to get this working now.
I don’t even know if this is an option but have you tried it in safe mode? You could possibly use MSConfig (start | run | msconfig ) to do a selective startup and turn things on/off to see if something is fighting with the IDE application’s timing requirements. I know I have used some debuggers that will not function properly is windows media player is even open. Difficult to find if it’s there but perhaps you have something like that going on.
Incorrect Eddie. The echo’d character is wired(physically the output from the PC is wired to the input AND output of the module. So when the PC sends something to the module it is also sending itself the same character. That is what a “Loop Back” echo is. Its the same as if the PCs output port was wired to it’s own input port. This means the PC did not see it’s own sent character. This means it’s a PC side problem(either hardware or software).
I suspect it is a software issue on the PC. If it was hardware I wouldn’t expect it to show up on the other devices he tried on that PC(eg USB to serial and PCI Serial adapter card). It is low level software though because the character was lost before the low level scanner could see it.
I’m familliar with half duplex. The loop back is at the atom end and it’s “possible” the atom crashed the byte. I am guessing you are assuming that because you got a handshake character back from the atom that it is on the host side that the byte disappeared. Is your IDE reporting or ignoring hard serial errors like framing? It is difficult to just make a byte disappear with no indications of where it went. An assumption there is that your application owns and holds the COM port for the duration of the transfer, although at the same time how wierd would it be for some other program to be sitting there jonesing to grab the serial port the instant another program let it go.
Anyway, personally I also think it is possibly some application or service, or maybe just even a crappy driver implemented for the motherboard chipset or i/o expander providing the COM port. Might be something you can find surfing forums or tech support FAQs for your mainboard or system part number Evolution… but a lot like looking for a needle in a haystack I think.
There isn’t any way for the AtomPro to “crash” the byte while in it reading loop. It is a half duplex, as you illuded to in your post, bit banged system which means the AtomPro is busy reading the bits being sent. Also if somehow the AtomPro got out of the reading loop(say because of a reset) to send data down and crash the byte being sent by the PC it would end up sending back a different ACK chacracter(0xAA). The 0xAA ACK chacrater is sent in the main command loop of the bootloader so if the AtomPro got reset you’d see that instead of the 0x55 ACKs. Also if it some how crashed the byte some other way I would expect a framing error or at least a garbage byte in the port monitoring program instead of no byte returned at all before the ACK character and no indication of any error at the port. That is why I’m pretty confident it’s software on the PC causing the problem. I guess it is possibly still a hardware issue(eg the AtomPro getting glitched somehow) but I’d have to say that I think the odds of that are getting up there with Mega Ball Lottery and Meteors hitting people heads.
I did try booting in safe mode and enabling the serial port. Same thing.
I have a blank 250gig backup drive I’m going to partition and install XP Pro on. I’ll try it in there and see if it works to eliminate the possibility that it’s the Media Center Edition of XP causing the problem. Then I’ll move on to drivers.
I could not tell for sure from one of your previous posts, but have you tried an USB to serial adapter on this machine as well and that this adapter worked on your portable? On My dev machine (this one), I am using a Linksys powered high speed USB hub with one of AcidTechs favorite USB to serial adapters and it is working good. If you have not tried this, you might try going down that avenue.
Kurt
Tried an IOGear USB to Serial and same result. Also tried the USB to serial on my laptop and it worked fine.
This is a long shot, but you might want to verify that your USB port on your PC provides power to your USB adapter. You can find this out by going to the device manager and finding which USB controller and USB HUB that your cable is connected to. If you look at the power page for those devices you should see something like: the hub provides 500mA per port and your device is using so much. In my case the Prolific one says it is using 100mA. I know on my portable and previous DEV machine I needed the external USB Hub to make mine work.
Kurt
WOW, MAJOR UPDATE. I figured it out. Well not really but I figured out how to make it work.
There were 2 things that were causing problems and after eliminating both, it works now!!!
First I started changing resources for my com port and testing different configurations. Some change but nothing major.
This is using just the serial port on the back of the computer, no adapter.
Then I checked the power mapping on my USB ports like kurte said.
Now I have 8 USB ports on the back of my machine and 4 on the front. Each port is configured with 500mah of available power. Now the front 4 USB ports wire into the motherboard but it turns out that those 4 SHARE power with 4 on the back. So When I checked the power settings, my gamepad controller was using 450mah plugged into the front and video capture box was using 500mAh. Which would explain why my video capture wasn’t working every now and then. Probably anytime my gamepad was plugged in. Different problem, also now solved.
So I disconnected every USB device from my system except the keyboard and mouse. All the power requirements were 0mAh except 100mah for the keyboard and mouse.
Didn’t think it could possibly be related but tried the Atom Pro IDE anyways. Bang, she programs all the way through past 100% and now hangs up at 50% verification. Tried about 10 times, always made it through programming, which it never did before, but hung up around the middle of verifying. NOW I’m getting somewhere.
Next, I looked at my Fifo buffers on the serial port more closely. Turns out that the 2 buffer sliders aren’t equal. Set on low they’re both on 1. but on high, the receive buffer is 14 and the transmit buffer is 16. now when you move the sliders to any one of the 3 preset “notches”, the transmit and receive values are always 1 off from each other.
So I slid the receive buffer to a value of 4. And instead of visually matching the bottom slider I moved it back a hair so it had a value of 4 instead of 5, where it wanted to snap to.
So now the sliders were off but the values matched.
Went into the Atom Pro IDE. Now my atom beeps when I click “Find Atom” Hit program and it scoots through programming and zips through verifying then bang, she’s done and the Atom give me a friendly beep to let me know at the end.
So I don’t understand how overloaded USB power can effect my Serial port or why the Fifo buffer sliders are not equal but if you set them equal it will work. I also tried with the Fifo buffers on 14 and 14, and 8 and 8. Works perfect. But at 1 and 1 or any other “snapped” positions (7 and 8, 12 and 14, etc.) it won’t work.
Hope this makes sense to you guys. Thanks for all the help guys. I learned alot about my USB and Serial setting through this and thanks to you guys.
IT WORKS!!!
IT WORKS!!!
IT WORKS!!!
IT WORKS!!!
USB to serial adapter also works now and so does the PCI Serial Card.
The only way I can see the power levels of USB affecting your regular comport is if it isn’t really a regular comport but a USB to serial on the motherboard. Interesting. Glad you got it and hopefully this will be helpfull for other out there(especially USB to serial adapter users).
Something funny. I started having a similar issue. The fix was the same. Seems my laptops two onboard USB ports no longer will supply power properly(I noticed they were acting strange the other day with simply jiggling the connector caused a reboot). Programming started failing randomly. I switched to a powered USB hub and no more problems. Just to be safe though I do set my FIFO buffers to match(the values not the sliders). Also a symptom others may see on underpowered USB to serial adapters is a slow down during programming. Programming is about half as fast on the underpowered USB ports(atleast before it failes).
I have a feeling my serial port is a usb to serial conversion done on the motherboard and the tech guys from Gateway had no idea what he was talking about.
The last time I talked to a tech at a big company like Gateway that knew anything about anything was in 1993. 
My first instinct was to sticky this thread. However it’s pretty long. I will try to wrap it up into a condensed version.
It sounds like the motherboard was made during the serial / usb changover. I’m really glad it’s working for you. 8)